Journal of General Virology (2008), 89, 2275–2279
Short Communication
DOI 10.1099/vir.0.83685-0
Genetic linkage among human cytomegalovirus glycoprotein N (gN) and gO genes, with evidence for recombination from congenitally and post-natally infected Japanese infants Hainian Yan,1,2 Shin Koyano,3 Yuhki Inami,1 Yumiko Yamamoto,1 Tatsuo Suzutani,4 Masashi Mizuguchi,2 Hiroshi Ushijima,2 Ichiro Kurane1 and Naoki Inoue1
Correspondence
1
Naoki Inoue
2
Department of Virology I, National Institute of Infectious Diseases, Tokyo, Japan Department of Developmental Medical Sciences, Institute of International Health, Graduate School of Medicine, University of Tokyo, Tokyo, Japan
[email protected]
3
Department of Pediatrics, Asahikawa Medical College, Hokkaido, Japan
4
Department of Microbiology, Fukushima Medical University, Fukushima, Japan
Received 23 December 2007 Accepted 15 May 2008
Investigation of sequence polymorphisms in the glycoprotein N (gN; gp4273), gO (gp4274) and gH (gp4275) genes of human cytomegalovirus (HCMV) strains collected from 63 Japanese children revealed that their gO genotype distribution differed slightly from that of Caucasian populations and that there was a significant linkage between the gN and gO genotypes. Linkage of these genotypes in strains obtained from Caucasian populations has been reported, so our similar findings in Japanese infants are consistent with this, and suggest generality of this linkage. Sequence analysis suggests that recombination between two strains of different linkage groups occurred approximately 200 bp upstream of the 39-end of the gO gene. Further studies are required to elucidate differences in biological characteristics among the linkage groups and the selective constraints that maintain the linkage.
Human cytomegalovirus (HCMV) infects most people during childhood without clinical symptoms; it is the major viral cause of birth defects and developmental abnormalities. It is also associated with significant morbidity and mortality in immunocompromised individuals. The complete genome of wild-type HCMV strains, such as Merlin, is 236 kb in length, and is predicted to encode 165 genes (Dolan et al., 2004). Genetic characterization of clinical isolates has mainly depended on sequence polymorphisms in the genes that encode viral envelope glycoproteins and cellular homologues (Rasmussen, 1999; Pignatelli et al., 2004). Glycoprotein B (gB), gM, gN, gH, gL and gO are involved in virus entry and egress and are the target molecules recognized by neutralizing antibodies. While extensive sequence variation is found in the gB and gH genes (5–10 %), a greater level is found in the gN and gO genes (40–50 %); the gL and gM genes are highly conserved among clinical strains. To date, the association of a particular genotype with a particular The sequence data determined in this study are available under GenBank accession numbers EU348337–EU348364. Two supplementary tables are available with the online version of this paper.
0008-3685 G 2008 SGM
Printed in Great Britain
clinical outcome has been controversial (Bale et al., 2000; Barbi et al., 2001; Trincado et al., 2000). We have recently investigated gB, UL144 and UL149 gene polymorphisms and found that the gB3 genotype was more prevalent in congenitally infected individuals with neurological abnormalities (Yan et al., 2008). Recently, a genetic linkage between the gN and gO genes was reported (Mattick et al., 2004). To learn how common this linkage is and whether the linkage group has any correlation with the clinical outcome of congenital infection, we analysed gN, gO and gH gene polymorphisms. HCMV strains were collected from 45 urine and 24 dried umbilical cord specimens obtained from 63 Japanese children, consisting of 32 congenitally and 31 post-natally infected children. Although samples of both materials were collected from six infants, specimens from each infant were handled as a single entity, as specimens from the same infant yielded the same sequence. Eleven of the congenital cases were identified previously by Ogawa et al. (2007). Six were identified by HCMV-specific IgM in maternal or cord blood specimens, a further six were identified by our HCMV screening programme (Inoue & Koyano, 2008) and 2275
H. Yan and others
the rest were identified by clinical manifestations. All congenital infections were confirmed by the detection of HCMV in urine within 2 weeks of birth or in dried umbilical cord specimens. Twenty-three healthy infants were chosen from .100 volunteers on the basis of the presence of HCMV in urine. HCMV was also collected from eight infants with hepatic damage, pneumonitis or bone marrow transplantation or infants that were born prematurely. Post-natal infection was implied by the absence of HCMV in their cord specimens. Viral DNA was extracted from these specimens as described previously (Ogawa et al., 2007). DNA fragments encoding hypervariable regions of the gN, gO and gH genes were amplified by nested PCR using Pfu polymerase (Promega) in 50 ml reaction volumes. Ten to 100 HCMV DNA copies were used as templates for the first-round PCR and 2 ml of these products were used for the second-round PCR. PCR conditions and primers are shown in Supplementary Table S1, available in JGV Online. The PCR products were separated on agarose gels and purified using a DNA extraction kit (QiaEX II, Qiagen). The purified DNA fragments were sequenced with BigDye Terminator Cycle Sequencing kit (Applied Biosystems) using the primers from the second-round of PCR. Sequences were assembled with ATGC version 4.0 (Genetyx, Tokyo) and aligned with Genetyx 7.0. Phylogenic analysis was performed with MEGA version 3.1 (Kumar et al., 2004). Designation of gN, gO and gH genotypes was based on previous publications (Mattick et al., 2004; Stanton et al., 2005; Chou, 1992; Pignatelli et al., 2003). The total number of available amplicons and the distribution of genotypes of congenitally and post-natally infected cases (including GenBank accesssion numbers) are given in Supplementary Table S2. There was no significant correlation between gH genotype and the incidence or clinical outcome of congenital infection. Since gO and gN have a large number of genotypes, the number of specimens in this study was insufficient to obtain statistically significant results in a clinical context. The presence of gO5 and the absence of gO1c in Japanese children made a slight difference in gO genotype distribution compared with the distribution in Caucasian populations (Mattick et al., 2004). Since the gO5 genotype has only recently been recognized (Stanton et al., 2005), the entire gO sequences of the five gO5 strains were determined. Their gO5 nucleotide sequences were identical to each other and were 99–100 % identical to those of Merlin and 3052. The gO genotypes exhibited a relatively low identity to each other; identity of the consensus gO5 sequence to other gO genotype sequences ranged from 76–81 % and 74–80 % at the nucleotide and the amino acid levels, respectively. Similar results were obtained from the phylogenic analyses for the sequences of the full-length and the middle segment of the gO gene (Fig. 1). Of the 63 analysed strains, 57 yielded a complete dataset for the gN, gO and gH genotypes (Table 1). gO and/or gH genotypes of the remaining six strains, one from a cord 2276
Fig. 1. (a) Phylogenic tree generated by the neighbour-joining method based on 60 nucleotide sequences covering the 440 bp middle part of the gO gene. Bootstrap values are indicated at the beginning of each major node. The frequencies of the gO genotypes among the 60 clinical strains are shown. Stars indicate the reference strains. (b) Phylogenic tree based on 13 full-length gO nucleotide sequences. Bar, 0.05 nucleotide substitutions per position.
specimen from a congenital case and five from urine specimens from healthy infants, were not available due to limiting amounts of HCMV DNA. Relationships were identified among the gN, gO and gH genotypes. For example, all gO1b strains link with gN3a and gH1, and all gO5 strains link with gN4c and gH2. Seven linkage groups cover 79 % (45/57) of the strains of these three genes for which sequences are available. If only the linkage between the gN and gO genotypes is considered, the seven groups cover 91 % (52/57) of the strains. Fisher’s exact test of the distributions of the matched genotypes between the gN and gO genes yielded a significant association (P,0.0001). Thus, our results clearly support the findings from a previous report (Mattick et al., 2004). As indicated in that study, breaking down the gO genotypes into seven or eight genotypes was critical to identifying the linkage that could not be previously identified in strains from US populations (Rasmussen et al., 2002, 2003). Although the relationship between gO/gN and gH genotypes seems to be significant, the small number of gH variations/genotypes limits the value of these statistical analyses. Two gH genotypes were not able to be divided into subgenotypes. The genotypes of gB do not correlate with gN, gO and gH genotypes (data not shown). In general, genetic polymorphisms and mosaic genotypes in the genome can be explained by the accumulation of spontaneous mutations under selective pressure and/or by homologous recombination, including both inter- and intra-strain recombination. Such recombination events have been reported for the variations within HCMV genes, such as gB (Haberland et al., 1999) and a duplicated pair of virokines (UL146 and UL147) (Arav-Boger et al., 2005). Recombination events have also been observed in other Journal of General Virology 89
HCMV gN-gO genotype linkage and interstrain recombination
Table 1. Linkage groups among the gN, gO and gH genotypes Strains were isolated from congenital (C) or post-natal (P) infections.
Strain
Infection type
ASA12 ASA59 C106 N42C C102 C140 C141 C177 FUK32 FUK72 FUK82 Y01 C83 J60250 U02 ASA16 FUK19U U01 U06 ASA68 C164 N66 C134 ASA15 FUK03 FUK20 FUK31 C145 J60236 U03 C49 J60249 J60248 ASA19 FUK74 X01 C110 C122 C14 C170 C185 C196 J60223 J60299 N22 C135 U07 ASA01 ASA70 FUK28 N59 C149
http://vir.sgmjournals.org
C C C C P P P C C C C C P P P C C P P C C C P C C C C P P P C P P C C C P P P P P P P P P P P C C C C P
Genotype gN gO gH 1 1 1 1 1 1 1 3a 3a 3a 3a 3a 3a 3a 3a 3b 3b 3b 3b 2 2 2 2 4a 4a 4a 4a 4a 4a 4a 4a 4a 4a 4b 4b 4b 4b 4b 4b 4b 4b 4b 4b 4b 4b 4b 4b 4c 4c 4c 4c 4c
1a 1a 1a 1a 1a 1a 1a 1b 1b 1b 1b 1b 1b 1b 1b 2a 2a 2a 2a 2b 2b 2b 2b 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2
Linkage group 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 3 3 3* 3* 4 4 4 4 5 5 5 5 5 5 5 5* 5* 5* 6 6 6 6 6 6 6 6 6 6 6 6 6* 6* 7 7 7 7 7
Table 1. cont.
Strain C154 FUK16 S01 J60284 J60298
Infection type C C C P P
Genotype gN gO gH 3a 3a 3a 3a 4c
1a 1a 1a 2b 1a
1 1 1 1 2
Linkage group D D D D D
*Strains with a gH genotype that is different to the others. DStrains that were not classified into the seven groups and could be products of recombination.
herpesviruses, including herpes simplex type 1 (Dutch et al., 1992), varicella-zoster virus (Norberg et al., 2006; Peters et al., 2006) and human herpesvirus 8 (Poole et al., 1999). Recombination depends on various immunological and intracellular constraints because infection of the host, ultimately of a single cell, with two parental strains is required. Concurrent infections with multiple HCMV strains have been observed in immunosuppressed patients, such as transplant recipients and patients with human immunodeficiency virus (Stanton et al., 2005; Puchhammer-Sto¨ckl & Go¨rzer, 2006; Coaquette et al., 2004). It has been demonstrated that pre-existing immunity does not prevent infection with strains of different genotypes (Ishibashi et al., 2007; Boppana et al., 2001). To find evidence indicative of recombination events, the entire gO gene sequences were determined for strains representing each genotype and for those that were not classified into the seven linkage groups. First, we analysed the gO5 strains. Whilst the gN4c genotype linked with the gO5 genotype in our population, the same gN4c genotype linked with the gO1c genotype in Caucasian populations. Mattick et al. (2004) discussed the possibility that gO1c was created by a recombination event. HCMV strains ASA01 and Toledo were chosen as representative strains of gN4c–gO5 and that of gN4c–gO1c, respectively. The similarity between the gN–gO sequences was analysed using the SimPlot program version 3.5 (http://asray.med. som.jhmi.edu/SCRofware/simplot) (Fig. 2a). The identity was .95 % from the gN gene to the 39-part of the gO gene, but it dropped significantly from 200 bp upstream of the 39end of the gO gene. This suggests that the gO5 strains are also the products of recombination. The recombination site could be anywhere within the conserved areas of the gN gene or the 200 bp 39-end region of the gO gene. If the unidentified counterpart for the recombination has a gN genotype other than gN4c, the transition site is expected to be around 200 bp upstream of the 39-end of the gO gene. Since gO1c is one of the rarest genotypes and no gO1c strain was identified in this study, further study is required to understand the relationship between the gO1c and gO5 strains. 2277
H. Yan and others
that of Y01, which represents linkage group 2 (gN3a– gO1b), and that of AD169, which represents linkage group 1 (gN1–gO1a). The sequence similarity of strain C154 with AD169 and Y01 declines around 200 bp upstream of the 39-end of the gO gene (Fig. 2b). This pattern suggests a potential recombination event between linkage groups 1 and 2 within the gO gene. The potential recombination site is assumed to be in the conserved 24 bp sequence shown in Fig. 2(b). Compared with the surrounding regions, the sequence around the recombination site did not necessarily correspond to the well-conserved regions (Mattick et al., 2004; Pignatelli et al., 2003). To visualize this, sequence similarity was compared for all eight linkage groups across the gN–gO locus (Fig. 2c). It is unlikely that this potential recombination event was due to the presence of two strains in the specimen or to mispriming, since (i) infection with multiple strains was not identified at detectable levels (.25 % in the population) in the raw sequence data of the gN, gO, gH, gB, UL144 and UL149 genes; (ii) three very similar but distinct strains were obtained from individuals from different localities and at different collection times and their DNAs were extracted and analysed in separate tests; and (iii) different primer sets yielded the same genotyping results. It is possible that the recombinant HCMV was generated and circulated naturally.
Fig. 2. Potential recombination events within the gO gene. The gN and gO genes are shown above the panels. The horizontal axis shows the nucleotide position starting from the 59-end of the gN open reading frame (a–c) and amino acid position in the gN and gO ORFs (d). (a) SimPlot analysis (window size 200, step size 20) of the similarity (%) between ASA01 (gN4c-gO5) and Toledo (gN4cgO1c). (b) SimPlot analysis (window size 250, step size 30) of C154 (gN3a-gO1a) with AD169 (gN1/gO1a; dashed line) and Y01 (gN3a-gO1b; dotted line). The divergent sequences around the potential recombination site are shown. (c) SimPlot analysis (window size 60, step size 1) of the mean similarity (identity score) among representative strains (AD169, Y01, FUK19, C164, C49, Towne, ASA1) of the eight linkage groups (seven groups described in this study and gN4c-gO1c). (d) VarPlot analysis (window size 20, step size 1) of the mean values of dN/ds ratio in the amino acid sequences of the gN and gO genes of the representative strains. dN/ds51 is indicated with a dashed line.
Next, we analysed three strains (C154, FUK16 and S01) of the gN3a–gO1a genotype, as they were not classified into the seven linkage groups. Since the gO sequences of these strains were almost identical (99–100 % identity), the gO sequence of C154 was used for further analysis. The sequence of the gN–gO locus of C154 was compared with 2278
Since only two cases of recombination were available, we could not tell whether the gO gene contains a hotspot sequence that triggers recombination, similar to those observed in other viruses (Magiorkinis et al., 2003; Kajino et al., 2001; Takeuchi et al., 2008). It was, however, confirmed that no chi site- or V(D)J recombination sitelike sequences were present in the gO and gN genes. To obtain insights into the mechanism of the recombination in the gO gene, the non-synonymous distance (dN) and synonymous distance (ds) of codon-based aligned gN and gO sequences were analysed using the VarPlot program, as a recent study provided evidence of positive selection in the hypervariable gN sequences (Pignatelli et al., 2003). In addition to the gN sequence, the gO sequence showed generally low dN and ds values, and the dN/ds ratios were almost all less than 1 (Fig. 2d), indicating that negative pressure tends to maintain the original sequences. Although dN/ds ratios .1 were observed in limited domains from some genotypes, such as the gO1b, gO2a, gO2b and gO4 sequences, the potential recombination sites do not localize at those positively selected sites, suggesting that positive selective pressure, such as for immune escape, is not providing selection for recombination. In conclusion, our study demonstrated a significant link between the gN and gO genotypes in Japanese infants, which supports a previous finding in Caucasian populations and suggests generality of the linkage. Whilst we describe a novel homologous recombination event in the gO gene, it will be important to identify additional recombination events in the gO gene in order to explain the mechanisms regulating recombination. Further studies are also required to elucidate differences in biological Journal of General Virology 89
HCMV gN-gO genotype linkage and interstrain recombination
characteristics among the linkage groups and to identify the selective constraints that maintain the linkage.
determination of hot spots and correlation with sequence similarity. J Gen Virol 84, 2715–2722. Mattick, C., Dewin, D., Polley, S., Sevilla-Reyes, E., Pignatelli, S., Rawlinson, W., Wilkinson, G., Dal Monte, P. & Gompels, U. A. (2004).
Acknowledgements We thank Yasushi Ohkusa, Phil Pellett, Naoki Nozawa and Kenji Fujieda for their intellectual input, and Risa Taniguchi and Hitomi Komura for their technical assistance. This study was supported by a grant (H20-Kodomo-007) from the Ministry of Health and Welfare (N. I.) and a fellowship from the Atsumi International Scholarship Foundation (H. Y.).
Linkage of human cytomegalovirus glycoprotein gO variant groups identified from worldwide clinical isolates with gN genotypes, implications for disease associations and evidence for N-terminal sites of positive selection. Virology 318, 582–597. Norberg, P., Liljeqvist, J. A., Bergstro¨m, T., Sammons, S., Schmid, D. S. & Loparev, V. N. (2006). Complete-genome phylogenetic
approach to varicella-zoster virus evolution: genetic divergence and evidence for recombination. J Virol 80, 9569–9576.
References
Ogawa, H., Suzutani, T., Baba, Y., Koyano, S., Nozawa, N., Ishibashi, K., Fujieda, K., Inoue, N. & Omori, K. (2007). Etiology of severe sensorineural
Arav-Boger, R., Zong, J. C. & Foster, C. B. (2005). Loss of linkage
hearing loss in children: independent impact of congenital cytomegalovirus infection and GJB2 mutations. J Infect Dis 195, 782–788.
disequilibrium and accelerated protein divergence in duplicated cytomegalovirus chemokine genes. Virus Genes 31, 65–72. Bale, J. F., Jr, Murph, J. R., Demmler, G. J., Dawson, J., Miller, J. E. & Petheram, S. J. (2000). Intrauterine cytomegalovirus infection and
glycoprotein B genotypes. J Infect Dis 182, 933–936. Barbi, M., Binda, S., Caroppo, S., Primache, V., Dido`, P., Guidotti, P., Corbetta, C. & Melotti, D. (2001). CMV gB genotypes and outcome of
vertical transmission: study on dried blood spots of congenitally infected babies. J Clin Virol 21, 75–79. Boppana, S. B., Rivera, L. B., Fowler, K. B., Mach, M. & Britt, W. J. (2001). Intrauterine transmission of cytomegalovirus to infants of
women with preconceptional immunity. N Engl J Med 344, 1366–1371. Chou, S. (1992). Molecular epidemiology of envelope glycoprotein H
of human cytomegalovirus. J Infect Dis 166, 604–607. Coaquette, A., Bourgeois, A., Dirand, C., Varin, A., Chen, W. & Herbein, G. (2004). Mixed cytomegalovirus glycoprotein B genotypes
in immunocompromised patients. Clin Infect Dis 39, 155–161. Dolan, A., Cunningham, C., Hector, R. D., Hassan-Walker, A. F., Lee, L., Addison, C., Dargan, D. J., McGeoch, D. J., Gatherer, D. & other authors (2004). Genetic content of wild-type human cytomegalovirus.
J Gen Virol 85, 1301–1312. Dutch, R. E., Bruckner, R. C., Mocarski, E. S. & Lehman, I. R. (1992).
Peters, G. A., Tyler, S. D., Grose, C., Severini, A., Gray, M. J., Upton, C. & Tipples, G. A. (2006). A full-genome phylogenetic analysis of
varicella-zoster virus reveals a novel origin of replication-based genotyping scheme and evidence of recombination between major circulating clades. J Virol 80, 9850–9860. Pignatelli, S., Dal Monte, P., Rossini, G., Chou, S., Gojobori, T., Hanada, K., Guo, J. J., Rawlinson, W., Britt, W. & other authors (2003).
Human cytomegalovirus glycoprotein N (gpUL73-gN) genomic variants: identification of a novel subgroup, geographical distribution and evidence of positive selective pressure. J Gen Virol 84, 647–655. Pignatelli, S., Dal Monte, P., Rossini, G. & Landini, M. P. (2004).
Genetic polymorphisms among human cytomegalovirus (HCMV) wild-type strains. Rev Med Virol 14, 383–410. Poole, L. J., Zong, J. C., Ciufo, D. M., Alcendor, D. J., Cannon, J. S., Ambinder, R., Orenstein, J. M., Reitz, M. S. & Hayward, G. S. (1999).
Comparison of genetic variability at multiple loci across the genomes of the major subtypes of Kaposi’s sarcoma-associated herpesvirus reveals evidence for recombination and for two distinct types of open reading frame K15 alleles at the right-hand end. J Virol 73, 6646–6660. Puchhammer-Sto¨ckl, E. & Go¨rzer, I. (2006). Cytomegalovirus and
Epstein-Barr virus subtypes–the search for clinical significance. J Clin Virol 36, 239–248.
Herpes simplex virus type 1 recombination: role of DNA replication and viral a sequences. J Virol 66, 277–285.
Rasmussen, L. (1999). Molecular pathogenesis of human cyto-
Haberland, M., Meyer-Ko¨nig, U. & Hufert, F. T. (1999). Variation
Rasmussen, L., Geissler, A., Cowan, C., Chase, A. & Winters, M. (2002). The genes encoding the gCIII complex of human cyto-
within the glycoprotein B gene of human cytomegalovirus is due to homologous recombination. J Gen Virol 80, 1495–1500.
megalovirus infection. Transpl Infect Dis 1, 127–134.
Inoue, N. & Koyano, S. (2008). Evaluation of screening tests for
megalovirus exist in highly diverse combinations in clinical isolates. J Virol 76, 10841–10848.
congenital cytomegalovirus infection. Pediatr Infect Dis J 27, 182–184.
Rasmussen, L., Geissler, A. & Winters, M. (2003). Inter- and
Ishibashi, K., Tokumoto, T., Tanabe, K., Shirakawa, H., Hashimoto, K., Kushida, N., Yanagida, T., Inoue, N., Yamaguchi, O. & other authors (2007). Association of the outcome of renal transplantation with
intragenic variations complicate the molecular epidemiology of human cytomegalovirus. J Infect Dis 187, 809–819.
antibody response to cytomegalovirus strain-specific glycoprotein H epitopes. Clin Infect Dis 45, 60–67. Kajino, K., Yamamoto, T., Hayashi, J., Umeda, T., Takahara, T. & Hino, O. (2001). Recombination hot spot of hepatitis B virus genome
binds to members of the HMG domain protein family and the Y box binding protein family; implication of these proteins in genomic instability. Intervirology 44, 311–316. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform 5, 150–163.
Kumar, S., Tamura, K. & Nei, M. (2004).
Magiorkinis, G., Paraskevis, D., Vandamme, A. M., Magiorkinis, E., Sypsa, V. & Hatzakis, A. (2003). In vivo characteristics of human
immunodeficiency
virus
http://vir.sgmjournals.org
type
1
intersubtype
recombination:
Stanton, R., Westmoreland, D., Fox, J. D., Davison, A. J. & Wilkinson, G. W. (2005). Stability of human cytomegalovirus genotypes in
persistently infected renal transplant recipients. J Med Virol 75, 42–46. Takeuchi, Y., Myers, R. & Danos, O. (2008). Recombination and
population mosaic of a multifunctional viral gene, adeno-associated virus cap. PLoS ONE 3, e1634. Trincado, D. E., Scott, G. M., White, P. A., Hunt, C., Rasmussen, L. & Rawlinson, W. D. (2000). Human cytomegalovirus strains associated
with congenital and perinatal infections. J Med Virol 61, 481–487. Yan, H., Koyano, S., Inami, Y., Yamamoto, Y., Suzutani, T., Mizuguchi, M., Ushijima, H., Kurane, I. & Inoue, N. (2008). Genetic
variations in the gB, UL144 and UL149 genes of human cytomegalovirus strains collected from congenitally and post-natally infected Japanese children. Arch Virol 153, 667–674.
2279
